Auto time stepping

[MF87]

Hello,

I am using the auto time stepping very often in FEAP. It works quite well for my problems, but now I have the problem that I would like to compute exactly until a specified time (in my example until time=5.0, see below). Is there a possibilty that I really end with my computation exactly at time=5.0? So far FEAP is always running until a time that is really close to 5.0, but less than it (e.g 4.9999547). I could set the end time to 5.01 or 5.1, but then the computation is stopped at times greater than 5.0 (e.g. 5.0025689). Is there any possibility to end up exactly at time=5.0 with auto time stepping? I need this in my computation to compare different simulations.

Thanks and best regards Marek


Example:

BATCh
  DT,,0.1
  AUTO,TIME,6,12,10
  LOOP,time,10000
    TIME,,5.0
    loop,iter,20
      tang,,1
    next,iter
  NEXT time
END

[Yaakov]

Hi,

I hope, the following concept can give an excatly end time:

in BATCH ... End we need to use both kinds of time stepping:

1. ) an auto time stepping
2. ) a fixed time

in 1. )  we should define a value of time nearly the goal time (e.g. 4.991)
after firt step, you can get the current time t' ( e.g. 4.995)

in 2.) using of the fixed time steps for the remaining time:
 
e.g, 0.005 = dt* 100 steps; dt = 0.00005;

if \delta t is very small, just one step ...

Cheers, Yaakov

P.S. to get the current time, you can use TIME command as a  trick to specify the value of time.

 

[Prof. R.L. Taylor]

We should be able to fix this.  Can you send a small problem that we can use to debug your problem?

[Prof. R.L. Taylor]

What version are you using?  I ran a small test problem and it stopped on the right time -- but I am sure it did no adjust the time step much

[MF87]

I am using feap 8.4.1.d . I tried to find a small debug problem with a feap element, but I did not find one. All the examples I ran with feap elements it stopped at the right time. In my example (single-edge notched specimen) in which I use a user-element with elastic isotropic damage, the problem is only occuring when I want to stop in the range of the sudden drop of the force-displacement curve (see attached picture). I did not manage to find an example with a feap element where such a sudden drop is happening (we would need e.g. softening plasticity). During this drop, the time increment is decreased enormously by feap, which seems to be in my point of view the reason, why the auto time stepping with stop at the prescribed point is not working.
I additionally attached the PlateHole example from the example manual of feap 8.4, in which such a sudden drop in the force-displacment curve should also occur, when a softening material model is used. Unfortunately, I did not manage to let the computation run (so far I used a hardening model, but this is also not running). There was always the error: *WARNING* NO CONVERGENCE

Greetings from Aachen and thanks in advance
MF87